I'm Chen Siyuan (陈思远). Currently work in Hangzhou city as an Algorithm Engineer devoting into LLM Inference and AI Infra. I obtained my Master degree (by research) at Nanyang Technological University, Singapore (NTU), supervised by Prof Mao Kezhi. My research field is Knowledge-Augmented Natural Language Processing. Before that I got my Bachelor degree from University of Electronic Science and Technology of China (UESTC), majoring in Digital Media Technology at School of Computer Science and Engineering.
Obtained my Bachelor degree from School of Computer Science and Engineering, University of Electronic Science and Technology of China (UESTC). Majoring in digital media technology.
2017~2021Obtained MEng degree advised by Prof Mao Kezhi at School of Electrical and Electronic Engineering, Nanyang Technological University (NTU). My research area is Knowledge-Augmented NLP and I explored the techniques of incorporating external knowledge into neural networks on NLP tasks such as sentiment analysis and relation extraction.
2021~2023Aug 2022 – Jan. 2023. Proposed to integrate both explicit and implicit knowledge with graph-based neural networks for event causality identification. Our model achieved the best results on 3 benchmark datasets. Published as Siyuan Chen and Kezhi Mao. 2024. Explicit and implicit knowledge-enhanced model for event causality identification. Expert Systems with Applications 238 (March 2024), 122039.
Spet. 2021 – Aug 2022. Proposed a novel GAT-based model which utilizes external knowledge and multi-granular information. Our model achieved the best results on the benchmark datasets. Published as Siyuan Chen and Kezhi Mao. "A graph attention network utilizing multi-granular information for emotion-cause pair extraction." Neurocomputing, vol. 543, p. 126252. 2023.
Jan. 2021 – June 2021. Proposed a novel embedding-based knowledge graph completion model that is able to deal with unseen entities and achieved better-than-baseline performance on DBPedia50k and FB20k datasets, and implemented a GUI system using PyQt.
Oct. 2019 – Jan. 2020. Applied Matching Networks to classify skin lesion images. Dataset is ISIC 2018 Task 3. After using data augmentation techniques such as translation and rotation, the accuracy reached 76%. Supervised by Prof Chen Juan.
Algorithm Engineer at Huawei company. Field of research is AI Infra, especially in LLM Inference. Hands on experience on LLM serving systems/frameworks such as SGLang and MindIE (closed-source LLM serving framework developed by Huawei Ascend). Familiar with core techniques in LLM serving systems, e.g. PD disaggregation, BatchScheduler, distributed kvcache.
Jan 2024 - Oct 2025Senior Algorithm Engineer at Ant International, focusing on LLM inference.
Oct 2025
